EN

Engineering · architecture · any industry

Development of systems of any complexity

My technical partner Sergey Shibut and his team lead this work. It is for projects that need a complete system rather than just a website, across industries and project sizes.

Looking for marketing and new customers? Return to Marketing

Specialities

What we build

Systems whose hardest work happens internally: data structures, access controls, integrations and what happens when user numbers grow tenfold.

What are these systems?

  • User accounts
  • Mobile applications
  • Loyalty programs
  • Dealer portals
  • Internal platforms and admin panels
  • Billing and settlements
  • Integration Gateways

If your task is not on the list, describe it in your own words. We will review it and tell you whether we can take it on.

What does such a system consist of?

The same six layers. The difference between a simple system and a complex one is how much attention each requires.

01data model

How the data is organized

The interface is easy to change. The underlying data structure is much harder. A mistake may become apparent a year later, after millions of records have accumulated and redesign is difficult. We therefore start by understanding your business: customers, orders, contracts and devices, how they relate, and which actions are allowed.

  • A data schema that will survive growth and changing rules
  • Migrations without stopping work and without losing history
  • The system keeps its data consistent, even during a failure
02access rights and activity log

Who sees what and who changed what

As soon as there is more than one type of user in a system, the main question becomes “who sees what.” Branches, contractors, partners, managers. Everyone has their own piece of data. Next to it is the question “who changed it”, without an answer to which the analysis of any incident comes down to guesswork.

  • Roles and access rights that reflect your company structure
  • Action log, which shows who changed what
  • Login by phone, second factor, session management
03integrations

What the system connects to

A system rarely operates alone. It connects to accounting, point-of-sale systems, telephony, payments, specialist software and equipment. Each can respond differently, fail differently or change its format without warning. We design integrations so an external failure does not bring your system down.

  • 1C, payment gateways, telephony, instant messengers
  • Industry-specific systems and equipment, from clinic information systems to inventory management
  • Scheduled and event-driven data exchange, with retries after failure
04load and fault tolerance

What will happen when there are more users?

We plan capacity before the first traffic spike. The difference between a system for a hundred users and one for a hundred thousand is not simply server power. It depends on where the bottleneck lies, usually the database, a queue or the slowest external service.

  • Caching, queues and background tasks. Heavy calculations run in the background so users do not have to wait
  • Database replicas and spare capacity for peak load
  • When overloaded, the system does not crash completely. Some functions are disabled, the rest works
05business logic and calculations

Who changes pricing and rules

Rewards, pricing plans, discounts, limits and settlements. Businesses change these rules most often, so we move them out of code and into settings. Otherwise, every promotion needs a software release and a developer to become available.

  • A marketer can change reward rules in the admin panel without a developer
  • Accurate calculations: money, rewards and balances
  • Reports generated automatically, without manual reconciliation in Excel
06operations and ongoing development

How the system is run after launch

A system is built once but runs for years. Over that time, support and improvements cost more than the original development. Code alone is therefore not enough. You need visibility into its operation and a way to hand it over to another team without lengthy investigation.

  • Monitoring and incident response under an agreed procedure
  • Backups and verified recovery
  • The source code, diagrams and documentation remain with you

Special case

If the system already exists

We do not always build from scratch. Clients also bring existing systems that have become too hard to develop. We review and replace them piece by piece while they keep working.

Problems clients bring to us

  • The developer left, and there is no one else to maintain the system
  • The contractor quit the job in the middle
  • Any edit breaks something nearby
  • The system cannot stop: the company relies on it every day

How we start

Understand the system before the first change

We run a copy, read the code and document how the system works today. Until that is done, every change is guesswork.

We identify what to keep and what to rewrite

Sometimes it's cheaper to fix something that exists than to write it all over again, and we'll say that.

We fix the most urgent problems first

We first fix failures that are stopping work today. The larger plan can wait.

Rewriting in parts

We replace the system piece by piece, enabling each part separately. It keeps working throughout.

We work within the constraints of the existing system

We will not insist on scrapping everything and starting again. We work with what you have.

Showing someone your system can feel awkward, particularly if it has not been tidied up for a long time. There is no need to be embarrassed; most systems have similar issues.

Tell us what is happening, even if you do not currently have a development budget. We will review it and suggest where to start. Sometimes a few tips are enough; sometimes we can help further.

Tell us about your system

Process

The work process

We first review the task and estimate the work, then write code. The estimate follows the review.

01

Understanding the task

We review what already works, where work is manual and where money is lost. The result is a description of the system and the scope of the first stage.

02

Prototype and estimate

We demonstrate the interface with a clickable prototype. Once it is agreed, we give a timescale and price. Before that, the figures would be guesswork.

03

Development

We work in short iterations. Every two weeks we show a working part of the system. The percentages of completion in the reports mean nothing.

04

Launch and handover

We migrate data, train staff and stay in touch. The source code and documentation belong to you.

Cooperation

Three work formats

We'll figure out which one is right for you during the first conversation.

Monthly

Retainer

Pay for a month of work

You pay a fixed monthly fee for the team to work on your system. Add tasks as you go and choose their priority.

Suitable for ongoing work on a system that keeps growing and needs continual changes.

The price is known in advance

Fixed amount

We fix the amount before the start

The task is described, the boundaries are clear. We name the amount and period before the start of work, then the price does not change.

Suitable for a specific improvement or a clearly defined piece of work.

From idea to launch

Entire project

We estimate each stage after reviewing the task

We take on a project when there is only an idea, before a solution or brief exists. We work out how it should function, design and develop it, launch it, then hand over the code and documentation.

Suitable when you need a finished result with minimal involvement from you.

Budget

$4,360

typical project budget

The price can be lower or higher. It depends on the task and the number of people needed.

A retainer usually costs less; a large project costs more. We give a price range straight away so you do not waste time if it is outside your budget. The exact price follows the task review and prototype.

Technologies

Our technology choices

We do not use a fixed set of tools. We choose them to fit the task. Below are the technologies the team already knows.

Languages

What is the system itself written in?

  • TypeScript
  • JavaScript
  • Python
  • Go
  • PHP
  • SQL

Backend and API

Logic, calculations and data exchange on the server side.

  • Node.js
  • NestJS
  • Express
  • FastAPI
  • Django
  • Laravel
  • REST
  • GraphQL
  • WebSocket
  • gRPC

Frontend

What the user sees and touches.

  • React
  • Next.js
  • Vue
  • Nuxt
  • Astro
  • Tailwind CSS
  • Vite

Mobile development

Applications for iPhone and Android.

  • Swift
  • SwiftUI
  • Kotlin
  • Jetpack Compose
  • React Native
  • Flutter

Data and storage

Where data is stored and how it is protected from loss.

  • PostgreSQL
  • MySQL
  • Redis
  • ClickHouse
  • MongoDB
  • Elasticsearch
  • S3

Queues and background tasks

The hard work moves out of the interface and into the background.

  • RabbitMQ
  • Kafka
  • Celery
  • BullMQ
  • cron

Infrastructure and Operations

Where the system runs, how it is updated and how it is monitored.

  • Docker
  • Kubernetes
  • nginx
  • CI/CD
  • Terraform
  • Prometheus
  • Grafana
  • Sentry

Team

Who is leading the development?

Sergey Shibut - Technical partner

Sergey Shibut

Technical partner

@Xenik

Sergey Shibut leads complex development. He writes code himself and assembles the right-sized team for each project, covering server-side systems, interfaces, mobile apps, integrations and testing.

His responsibility is the part that is hardest to change later: data structures, access rights and system behaviour under load. These early decisions shape everything that follows.

Rubber duck debugging

Sergey has a yellow rubber duck on his desk. Explain a problem to it aloud, and the solution can emerge halfway through. That is why our work starts with a conversation.

Our experience

  • Nationwide loyalty programmes used every day
  • User accounts and dealer portals for large networks
  • High-traffic services that continue to operate and evolve

How to demonstrate work under NDA

All of the team’s projects are covered by non-disclosure agreements, so company names are not disclosed here.

We can still demonstrate the work. During a call, we run a local copy and explain its data structures, access controls and behaviour under load.

If necessary, we will sign the same agreement with you.

Contacts

Tell us what needs to be built

Describe the task in your own words. Sergey leads development, joins the call and estimates the timescale.

or
Message me on Telegram